Imputation of gross amounts from net incomes in household surveys: An application using EUROMOD

Household micro-datasets often do not contain information on gross incomes. We present an algorithm which exploits the tax- and contribution rules built into tax-benefit models to convert net income information into gross amounts. Using EUROMOD, a multi-country taxbenefit model covering all fifteen countries of the European Union, net-to-gross conversions can be performed for a large number of countries utilising existing models of relevant fiscal rules. The algorithm takes into account all relevant complexities of tax- and contribution rules and can, thus, produce much more accurate results than statistical models which estimate netto-gross ratios using only a few explanatory variables. Among the features of the algorithm is the ability to distinguish between different individuals in the same household. Even if individuals’ incomes are taxed jointly, the algorithm is able to approximate separate net-togross factors for individuals in the same fiscal unit. This is possible since EUROMOD can accurately assign people to appropriate fiscal units. In addition, it is in certain cases possible to produce different net-to-gross ratios for different income components. We undertake a case study to illustrate the importance of deriving separate net-to-gross factors for different individuals within a household/fiscal unit and for different income sources of the same individual.